4.14-stable review patch. If anyone has any objections, please let me know.
------------------
From: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
[ Upstream commit 993b9bc5c47fda86f8ab4e53d68c6fea5ff2764a ]
The commit ca876c7483b6
("gpiolib-acpi: make sure we trigger edge events at least once on boot")
added a initial value check for pin which is about to be locked as IRQ.
Unfortunately, not all GPIO drivers can do that atomically. Thus,
switch to cansleep version of the call. Otherwise we have a warning:
...
WARNING: CPU: 2 PID: 1408 at drivers/gpio/gpiolib.c:2883 gpiod_get_value+0x46/0x50
...
RIP: 0010:gpiod_get_value+0x46/0x50
...
The change tested on Intel Broxton with Whiskey Cove PMIC GPIO controller.
Fixes: ca876c7483b6 ("gpiolib-acpi: make sure we trigger edge events at least once on boot")
Signed-off-by: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
Cc: Hans de Goede <hdegoede@xxxxxxxxxx>
Cc: Benjamin Tissoires <benjamin.tissoires@xxxxxxxxxx>
Acked-by: Mika Westerberg <mika.westerberg@xxxxxxxxxxxxxxx>
Signed-off-by: Linus Walleij <linus.walleij@xxxxxxxxxx>
Signed-off-by: Sasha Levin <alexander.levin@xxxxxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
drivers/gpio/gpiolib-acpi.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
--- a/drivers/gpio/gpiolib-acpi.c
+++ b/drivers/gpio/gpiolib-acpi.c
@@ -247,7 +247,7 @@ static acpi_status acpi_gpiochip_request
gpiod_direction_input(desc);
- value = gpiod_get_value(desc);
+ value = gpiod_get_value_cansleep(desc);
ret = gpiochip_lock_as_irq(chip, pin);
if (ret) {
